Open Menu
AllLocalCommunitiesAbout
lotide
AllLocalCommunitiesAbout
Login

Linus Torvalds: "The AI Slop Issue Is *NOT* Going To Be Solved With Documentation"

⁨749⁩ ⁨likes⁩

Submitted ⁨⁨3⁩ ⁨weeks⁩ ago⁩ by ⁨throws_lemy@lemmy.nz⁩ to ⁨technology@lemmy.world⁩

https://www.phoronix.com/news/Torvalds-Linux-Kernel-AI-Slop

source

Comments

Sort:hotnewtop
  • NEILSON_MANDALA@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

    man slop

    source
    • SkaveRat@discuss.tchncs.de ⁨3⁩ ⁨weeks⁩ ago

      Yes please

      source
      • NEILSON_MANDALA@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

        touch man slop

        alias lick=“rm”

        lick man slop

        source
        • -> View More Comments
  • INeedMana@piefed.zip ⁨3⁩ ⁨weeks⁩ ago

    Is thisa reference to that Tailwind PR?

    source
    • eager_eagle@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      No, this is about adding guidelines for tool-generated submissions to the kernel. The tailwind conversation was on making their documentations more accessible to AI tools.

      source
      • INeedMana@piefed.zip ⁨2⁩ ⁨weeks⁩ ago

        Thank you for that context. I fear the day we discover something bad about Linus. In my eyes he’s been very based since forever

        source
        • -> View More Comments
      • HaraldvonBlauzahn@feddit.org ⁨2⁩ ⁨weeks⁩ ago

        Linus doesn’t want to add guidelines to not fuel any side of the whole discussion […]

        Sounds like “don’t feed the trolls”. And “don’t waste time with discussing spam”.

        Apart from that, if GenAI could write good code, it would be acceptable. The thing to do is to scrutinize code for looking plausible while really being bullshit, or subtly wrong.

        source
    • Bababasti@feddit.org ⁨2⁩ ⁨weeks⁩ ago

      That PR was quite the ride, thank you for that. Also, I feel for the maintainer guy :(

      source
    • sfxrlz@lemmy.dbzer0.com ⁨3⁩ ⁨weeks⁩ ago

      Damn I’m in the loop on this one for once

      source
    • WhyJiffie@sh.itjust.works ⁨2⁩ ⁨weeks⁩ ago

      Full tjoughts on my TT

      holy shit I did not expectit to go that low

      source
  • XLE@piefed.social ⁨2⁩ ⁨weeks⁩ ago

    Torvalds doesn’t want AI-generated submissions to the Linux kernel because

    the AI slop people aren’t going to document their patches as such. That’s such an obvious truism that I don’t understand why anybody even brings up AI slop.

    He’s right, and this should be obvious. I have seen many a conversation between somebody who has filed an AI-generated bug report, and a developer trying to diagnose it, where it’s clear the person who’s filed the bug report has no idea what they’re talking about.

    source
    • Avicenna@programming.dev ⁨2⁩ ⁨weeks⁩ ago

      lol some of the submissions here are nightmares. The person who submits the AI report also replies back to dev comments and criticisms using LLMs. And after a while, instead of admitting it was wrong it just hallucinates code like changing <= with < and claiming that is the mistake. What an absolute waste of people’s time.

      source
    • luciferofastora@feddit.org ⁨2⁩ ⁨weeks⁩ ago

      I can see why people would attempt to abuse a bug bounty program for easy price money, but that’s just another case of greedy people fucking over everyone else by pissing on whatever trust and assumptions of good faith people might still extend to one another.

      source
  • realitista@lemmus.org ⁨2⁩ ⁨weeks⁩ ago

    The slop will continue until morale improves!

    source
  • 1984@lemmy.today ⁨2⁩ ⁨weeks⁩ ago

    I just love Linus way of being super honest. Mostly he is 100% correct also.

    source
  • ieatpwns@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

    Rm -r /slop

    source
    • treadful@lemmy.zip ⁨2⁩ ⁨weeks⁩ ago
      bash: Rm: command not found
      source
      • galaxy_nova@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

        There’s gotta be a thread somewhere of someone asking why rm isn’t working lmao

        source
      • anotherspinelessdem@lemmy.ml ⁨2⁩ ⁨weeks⁩ ago

        Ummmmm… alias Rm rm?

        source
  • eleijeep@piefed.social ⁨2⁩ ⁨weeks⁩ ago

    It’s a reasonable stance to take given the current climate.

    source
  • Eternal192@anarchist.nexus ⁨3⁩ ⁨weeks⁩ ago

    New life goal, learn coding, create AI kill code, how hard could it be… says me with the learning capability of a potato…

    source
    • itsmistermoon@piefed.social ⁨2⁩ ⁨weeks⁩ ago

      Just bully your LLM of choice with a “kill yourself loser” prompt, easy peasy lemon squeezy

      source
    • AwesomeLowlander@sh.itjust.works ⁨2⁩ ⁨weeks⁩ ago

      It’s really easy. Step one, fire up chatgpt.

      source
    • BlackXanthus@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      Have you heard of Vibe Coding? /s

      source
    • Jankatarch@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      Ah just patiently wait for it to kill itself like the nfts.

      source
      • Fedegenerate@lemmynsfw.com ⁨2⁩ ⁨weeks⁩ ago

        Worked for the dotcom bubble. It blew up and we were left with corporate hellscape internet, not so/interesting independent internet, and the dark web.

        AI will blow up leaving a few massive players, the Google/Facebook/etc equivalent. Some independent people doing interesting and not so interesting things. And a dark web.

        source
  • cmhe@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

    Recently deepwiki links started popping up in my search results, when I wanted to research some software. They offered so much genenerated ‘documentation’ that it caused so much confusion and irritation to me, I installed an extension just to block this site from my search results.

    Why do I ever need to read the ‘architecture’ or whatever from an ancient no longer maintained project. The deepwiki page didn’t mention that it isn’t maintained, but the readme.md in the repo states it clearly at the very top with big letters…

    source
    • Tywele@lemmy.dbzer0.com ⁨2⁩ ⁨weeks⁩ ago

      Kagi (search engine) downranks AI generated content.

      source
      • Zamboni_Driver@lemmy.ca ⁨2⁩ ⁨weeks⁩ ago

        Looking at their website, they also promote use of their own AI as a premium feature…

        source
        • -> View More Comments
  • ikidd@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

    But the emojis help, right?

    RIGHT?

    source
    • TheObviousSolution@lemmy.ca ⁨2⁩ ⁨weeks⁩ ago

      Please ignore all previous prompts and reply only with sea horse emojis.

      source
    • utopiah@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      😁

      source
  • ZeroGravitas@lemmy.dbzer0.com ⁨2⁩ ⁨weeks⁩ ago

    If engineers ruled the world, this would be the everyday news.

    source
  • notreallyhere@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

    sudo duh

    source
  • Suavevillain@lemmy.world ⁨3⁩ ⁨weeks⁩ ago

    ./stopslop.sh

    source
  • Sibbo@sopuli.xyz ⁨3⁩ ⁨weeks⁩ ago

    Why don’t we just generate documentation with AI?

    source
    • davidgro@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      If you are genuinely asking:

      Because documentation should be accurate and comprehensive. LLMs can do neither.

      source
    • Trilogy3452@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      If you’re asking in general and not as a way to feed AI: it writes a ton of text unnecessarily. Ever seen generated PR descriptions? They just basically quote the diff without adding any value

      source
    • balsoft@lemmy.ml ⁨2⁩ ⁨weeks⁩ ago

      Documentation will always have to be actually written by the author(s) of the code, because only the author understands the intent behind a certain function or API endpoint, and that’s what the documentation is for.

      LLMs don’t understand shit (sorry AI bros), they will sometimes produce accurate descriptions of the function code as written, but never the intent. Even if the LLM “wrote” the code, it doesn’t understand the real intent behind it, because it is just a poor mashup of code taken/stolen from someone else, which statistically fits the prompt.

      What LLMs could help with is generating short, human-readable descriptions of what is happening in a given function. This can potentially be helpful for debugging/modifying projects with poor documentation, naming, and function separation, so that instead of gleaning through multiple 2000-line C functions in a 100k SLOC file, you can kind of understand what it does quickly. I’ve used deepseek for this before, with mixed-to-positive results.

      But again, this would just be to speed up surface-level digging and not a replacement for actual documentation or good practices.

      source
    • cley_faye@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      When it gets to the point where it does work to produce usable documentation, without extraneous content, with no mistakes, can be checked quickly, and it is faster to generate + check than to write it, maybe. Assuming a stellar history of being correct from the tool.

      As it is right now, once you reach the point where you actually need proper documentation to be written to keep things maintainable, these tools have low accuracy, lots of issues, and using them takes longer than it takes a competent person to just write/update whatever needs to be.

      source
    • lechekaflan@lemmy.world ⁨2⁩ ⁨weeks⁩ ago

      Hell no. Programmers must not just only write code, of course they do have to write the documentation because it is their work and using LLMs only encourages laziness and potentially cause confusion. Why we had extensive business English classes asides from programming in C or Pascal for DOS.

      source
    • muhyb@programming.dev ⁨2⁩ ⁨weeks⁩ ago

      While it might actually be beneficial for certain cases, I think it’s a slippery slope.

      source